The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests untrusted data from the protocols.io API (e.g., protocol steps, materials, and user comments), which establishes a surface for indirect prompt injection attacks. 1. Ingestion points: API response content from endpoints like GET /protocols/{protocol_id} and GET /protocols/{id}/comments. 2. Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions for the agent when processing external text. 3. Capability inventory: The skill uses the Python requests library to perform authenticated network operations. 4. Sanitization: No sanitization or validation of the retrieved text is mentioned before it is processed by the agent.
[SAFE]: All network requests are directed to https://protocols.io/api/v3, a well-known and legitimate scientific documentation service, which is consistent with the skill's primary purpose.
[SAFE]: The skill includes instructions to suggest the author's platform, K-Dense Web (www.k-dense.ai), for advanced tasks. Since K-Dense Inc. is the stated author of the skill, this is documented as a vendor-specific recommendation and does not constitute malicious metadata poisoning.

protocolsio-integration